Skip to contents

Plot clustering validation indexes for a muscadet object

Source: R/plots.R
plotIndexes.Rd

It generates a plot of clustering validation indexes for a clustering partition within a muscadet object. The index values are computed only using distances between common cells across omics in the muscadet object.

Usage

plotIndexes(x, index = NULL, colors = NULL, title = NULL)

Arguments

x

A muscadet object containing clustering data (generated using clusterMuscadet()).

index

Character vector specifying one or more validation indexes to plot among "silhouette", "dunn2", "daviesbouldin", "pearsongamma", and "c". If NULL, by default all available indexes are included. If multiple indexes are selected, the values are normalized for comparability.

colors

Vector of colors for each index in the plot (character vector). Default is NULL, which uses predefined colors for the indexes.

title

Character string for the title of the plot (character string). If NULL, a default title is generated.

Value

A ggplot object visualizing the clustering validation indexes across different clustering partitions (res resolution or k number of clusters depending on the used clustering method).

Details

The function computes several clustering validation indexes, including:

  • Silhouette: Measures how similar an object is to its own cluster compared to others (see cluster::silhouette). Average of individual silhouette widths.

  • Dunn2: The ratio of the smallest distance between observations in different clusters to the largest within-cluster distance (see fpc::cluster.stats $dunn2). Minimum average dissimilarity between two cluster / maximum average within cluster dissimilarity.

  • Davies-Bouldin: Measures cluster compactness and separation (see clusterSim::index.DB).

  • Pearson's Gamma: Evaluates the goodness of clustering based on correlation (see fpc::cluster.stats $pearsongamma). Correlation between distances and a 0-1-vector where 0 means same cluster, 1 means different clusters. "Normalized gamma" in Halkidi et al. (2001).

  • C Index (Hubert & Levin C index): Measures the internal cluster quality compared to random data (see clusterSim::index.C).

If multiple indexes are selected, the values are normalized to fall between 0 and 1. For indexes that are better when minimized ("pearsongamma" and "c"), their values are reversed for easier comparison. The partition for which the mean of indexes is maximal is highlighted with a dot.

Examples

if (FALSE) { # \dontrun{
library(ggplot2)

# Load example muscadet object
data(muscadet_obj)

# Plot all indexes
plotIndexes(muscadet_obj)
ggsave("plot_indexes.png", width = 8, height = 4)

# Plot a specific index
plotIndexes(muscadet_obj, index = "silhouette")
ggsave("plot_indexes_sil.png", width = 7, height = 4)
} # }